A Hybrid Approach for Creating Knowledge Graphs: Recognizing Emerging Technologies in Dutch Companies

From text to model: Leveraging natural language processing for system dynamics model development

“…The NLP field offers starting points for such an endeavour. Bakker et al (2022a, 2022b) and Sil et al . (2010) have developed methods to extract combinations of actions, actors, objects, recipients, preconditions and postconditions from texts.…”

“…The NLP field offers starting points for such an endeavour. Bakker et al (2022aBakker et al ( , 2022b and Sil et al (2010) have developed methods to extract combinations of actions, actors, objects, recipients, preconditions and postconditions from texts. The output of these methods can benefit operational thinking during SD modelling by identifying actors, their decisions, and subsequent actions and results, alongside the information relations that guide these processes.…”

“…Concepts and relations in a knowledge graph can be input from one NLP model or a collection of NLP models, each focused on extracting specific concepts or semantic relations. Hybrid forms of knowledge graph creation are being developed that allow users to evaluate and direct the graph construction process (Bakker et al, 2022a(Bakker et al, , 2023de Boer & Verhoosel, 2020;Opasjumruskit et al, 2022). This could enable a modeller to guide the process based on the purpose of the SD model and other considerations-for instance, by indicating which concepts in the graph are relevant, irrelevant or can be aggregated.…”